首页 文章

Keras是否具有将输入字向量复制并仅反向传播到一组中的功能？

提问于
浏览 次
2

我的目标是在Keras创建一个RNN-CNN网络，根据文本段落预测分类输出 . 在我当前的模型中，段落首先被嵌入到特征向量中，馈送到2个cuDNNGRU层，4个Conv1D和MaxPooling层，然后到达密集输出层 .

但是，我发现了一种多通道方法的参考，用于处理涉及复制初始向量的单词向量，通过CNN层运行一组，然后在汇集之前将输出与副本相加 . 这样做是为了防止反向传播到一组向量中，因此保留了原始单词向量中的一些语义概念 .

我已经尝试过搜索这个但是与多声道和CNN相关的唯一事情就是使用多种大小的n-gram内核 . Keras是否提供可用于实现此目的的任何功能？

python keras nlp deep-learning conv-neural-network

1 回答

  • 1

    是的，您可以通过使用函数API来实现此目的 .

    这是一个小例子随时可以适应您的需求：

    embed_input = Input(shape=(300,))
embedded_sequences = Embedding(10000, 10)(embed_input)
embed=SpatialDropout1D(0.5)(embedded_sequences)

gru=Bidirectional(CuDNNGRU(200, return_sequences = True))(embed)

conv=Conv1D(filters=4,
            padding = "valid",
            kernel_size=4,
            kernel_initializer='he_uniform',
            activation='relu')(gru)

avg_pool = GlobalAveragePooling1D()(conv)
max_pool = GlobalMaxPooling1D()(conv)
gru_pool = GlobalAveragePooling1D()(gru)


l_merge = concatenate([avg_pool, max_pool, gru_pool])



output = Dense(6, activation='sigmoid')(l_merge)
model = Model(embed_input, output)


model.summary()


output:

__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
input_10 (InputLayer)           (None, 300)          0                                            
__________________________________________________________________________________________________
embedding_9 (Embedding)         (None, 300, 10)      100000      input_10[0][0]                   
__________________________________________________________________________________________________
spatial_dropout1d_9 (SpatialDro (None, 300, 10)      0           embedding_9[0][0]                
__________________________________________________________________________________________________
bidirectional_8 (Bidirectional) (None, 300, 400)     254400      spatial_dropout1d_9[0][0]        
__________________________________________________________________________________________________
conv1d_6 (Conv1D)               (None, 297, 4)       6404        bidirectional_8[0][0]            
__________________________________________________________________________________________________
global_average_pooling1d_6 (Glo (None, 4)            0           conv1d_6[0][0]                   
__________________________________________________________________________________________________
global_max_pooling1d_6 (GlobalM (None, 4)            0           conv1d_6[0][0]                   
__________________________________________________________________________________________________
global_average_pooling1d_7 (Glo (None, 400)          0           bidirectional_8[0][0]            
__________________________________________________________________________________________________
concatenate_5 (Concatenate)     (None, 408)          0           global_average_pooling1d_6[0][0] 
                                                                 global_max_pooling1d_6[0][0]     
                                                                 global_average_pooling1d_7[0][0] 
__________________________________________________________________________________________________
dense_5 (Dense)                 (None, 6)            2454        concatenate_5[0][0]              
==================================================================================================
Total params: 363,258
Trainable params: 363,258
Non-trainable params: 0

    和图的结构：

    enter image description here

    回复于

相关问题